The Peak-Performance Analysis Method for Optimizing Any GPU Workload [wayback-archive]
- how to use GPU counters to detect interactions between different hardware blocks
- discussion of case studies
- walkthrough of analysis, code examples to show optimizations and discussion of results
Pre-Multiplied Alpha [wayback-archive]
- reduces color bleeding in alpha blending
- can represent Additive, Alpha Blend, and Multiply blend modes
- can vary the blend mode per-pixel, can be used to pre-combine effects
SRV handling in D3D12 - twitter thread [wayback-archive]
- how to design for descriptor tables in D3D12
- goal is to only set a few root constants to point to the correct data
Vulkan ASKs : Swapchain [wayback-archive]
- proposed vulkan extensions to improve swapchain and present interactions
Profiling: Measurement and analysis [wayback-archive]
- overview of profiler types
- visualization of the memory layout and how it causes performance problems
- how to detect cache misses with CodeXL
HLSLexplorer [wayback-archive]
- small tool that takes HLSL code, passes it to the shader compiler and shows the d3d11 shader disassembler output
- allows to select compiler flags and see the effects on the assembly
- similar to Pyramid, allows to see the native GPU disassembly as well
A new microflake model with microscopic self-shadowing for accurate volume downsampling [wayback-archive]
- technique to preserve anisotropic attenuation and self shadowing when downsampling volume representations
- allows silhouettes of LODs to be better preserved
Exploring the limits of real time rendering
- showcase of a photogrammetry nature scene render in UE4
